The information sensitivity of debt in good and bad times

Endogenous debt maturity and rollover risk

“…The residual category, NotExposed t:t+5 , equals one when both Exposed t and Exposed t+1:t+5 are equal to zero.22 In untabulated regressions, we allow for a continuous interaction between market expectations and the amount of expiring debt, obtaining qualitatively similar results.23 In order to keep a consistent measure of expiring debt across the two time periods,Table B2only shows results for the amount of expiring debt with original maturity of more than five years. Our findings are broadly consistent if, instead, we employ the amount of expiring debt issued before August 2007 for the crisis period and the total stock of expiring obligations in normal times.24 SeeDang, Gorton, and Holmstrom (2013) for a definition of information sensitivity andBrancati and Macchiavelli (2018) for evidence that debt is not informationally sensitive in the precrisis period.…”

“…The second source of endogeneity is due to market expectations: innovations in default risk can be observed by analysts and internalized in their forecasts; we then need to treat as an endogenous variable. To deal with this issue, we use the IV approach developed in Brancati and Macchiavelli (), which relies on Bayesian learning theory to recover a set of external instruments, namely past forecast errors. In particular, if market participants are uncertain about the dynamics of banks' fundamentals, Bayesian‐learning agents take advantage of past forecast errors to correct and update current expectations (for more details, see Brancati & Macchiavelli, ).…”

“…To deal with this issue, we use the IV approach developed in Brancati and Macchiavelli (), which relies on Bayesian learning theory to recover a set of external instruments, namely past forecast errors. In particular, if market participants are uncertain about the dynamics of banks' fundamentals, Bayesian‐learning agents take advantage of past forecast errors to correct and update current expectations (for more details, see Brancati & Macchiavelli, ). The construction of forecast errors is detailed in Appendix A.…”

“…Furthermore, not having information on the amount of expiring debt with original maturity of less than one year, if anything, biases the coefficient of interest ( 1 ) to correct and update current expectations (for more details, see Brancati & Macchiavelli, 2018). The construction of forecast errors is detailed in Appendix A.…”

“…See Dang, Gorton, and Holmstrom () for a definition of information sensitivity and Brancati and Macchiavelli () for evidence that debt is not informationally sensitive in the precrisis period.…”